On May 20, two highly synchronized events occurred in the Chinese enterprise software industry: Kingdee released the enterprise-level AI operating system "Lingji" in Shenzhen, while Yonyou's YonClaw exclusive invitation code also opened on the same day. This is no coincidence.
On the same day, China's two major ERP giants revealed their AI trump cards. The competition behind this is not about whose model is stronger, but about who has the right to define the standards of the enterprise AI era—namely, when enterprises procure AI capabilities in the future, whether they follow Kingdee's approach or take Yonyou's path.
Two Trump Cards: The Differences Between Spirit Base and YonClaw
First, let's lay both cards on the table. There is a fundamental difference in positioning between the two; they are not competing on the same dimension.
| Comparison Dimensions | Kingdee "Lingji" | Yonyou「YonClaw」 |
|---|---|---|
| Core Positioning | Enterprise-grade AI Operating System | Enterprise Super Agent |
| Design Concept | AI-native, designed from scratch with a brand-new paradigm | Native integration, embedding AI into the existing BIP system |
| Core Competencies | A complete technology stack for agent development + orchestration + operations + governance; finance-centric, connecting the entire enterprise in real time | Understand the task goal → Plan the execution path → Mobilize skills and systems to complete the entire process from asking questions to execution. |
| Adapt to Customers | Growing and mid-sized enterprises; startups that have never used ERP | Large conglomerates, central and state-owned enterprises (customers that have deployed Yonyou BIP) |
| Safety Certification | Passed the relevant standard tests of the China Academy of Information and Communications Technology; proposed a "deep security architecture" for the Agent era. | Passed the “Enterprise-Level Claw-Like Agent Security Capability Assessment Certification” of the China Academy of Information and Communications Technology |
One-sentence summary: Kingdee is building the “next-generation ERP,” while Yonyou is developing the “AI upgrade package for this generation of ERP.” These two paths are not superior or inferior to each other, but their customer bases and replacement costs differ by an order of magnitude.
Why Now: The Dual Window of Domestic Substitution and AI Transformation
To understand this move, one must consider the overlap of two time windows:
- Domestic substitution has entered the deep-water zone. Yonyou's public data shows that it has replaced SAP and Oracle systems for more than 270 central SOEs and state-owned enterprises. The significance lies in this: once SAP is replaced, subsequent AI upgrades are highly likely to continue using Yonyou systems, because migration costs are too high.
- The battle for the right to define the AI operating system has already begun.SAP unveiled its "Autonomous Enterprise" vision at its 2026 Sapphire conference, but implementing it in the Chinese market requires localization. Both Kingdee and Yonyou have seen this gap—whoever defines the standard for enterprise AI operating systems first will be able to continue benefiting in the second phase of domestic substitution (AI upgrade).
Kingdee chose to release Lingji on May 20, with the timing just one day ahead of the national AI agent evaluation standard submitted for review on May 21. Those in the know naturally understand this rhythm.
What This Means for Customers
If you are the chief information officer of a manufacturing enterprise, you now face roughly the following three options:
Option 1: Wait for SAP/Oracle's AI versions to be deployed in China. The timeline is uncertain, compliance risk is uncertain, and the price may not be cheap.
Option 2: Adopt Kingdee Lingji. Suitable for enterprises that have not yet deeply integrated with a specific ERP system, or for greenfield deployment scenarios involving newly established subsidiaries or factories. Its AI-native architecture is theoretically more flexible, but ecosystem maturity still requires verification.
Option 3: Launch YonClaw on the basis of Yonyou BIP. Suitable for central state-owned enterprises and large groups that already use Yonyou products. There is no need to comprehensively overhaul existing systems, but the upper limit of AI capabilities will be constrained by the openness of the current architecture.
There is a practical issue that few people mention: switching costs in the ERP market are extremely high; once chosen, you are locked in for at least 5–8 years. Therefore, Kingdee is desperately trying to seize share in the large-enterprise market, while Yonyou is desperately trying to defend the central state-owned enterprises it has already won—the essence of this war is who will have the say over Chinese enterprises' AI infrastructure in the next five years.
Odoo's positioning in this field
Speaking of this, let me also mention Odoo by the way. The main subject discussed above is heavy-duty ERP for large and medium-sized enterprises, but China still has a large number of SMEs—they either use Kingdee Jingdouyun, Yonyou Chanjet, or are still using Excel.
Odoo's opportunity lies in this: The battleground for Lingji and YonClaw is medium- and large-sized enterprises, while the AI+ERP needs of small and medium-sized enterprises have not yet been seriously addressed by anyone. Odoo's open-source nature and modular architecture theoretically make it better suited to rapidly integrating various large models and achieving lightweight AI enhancements—such as automatically generating sales quotations, intelligently matching suppliers, and predicting inventory replenishment points. These scenarios do not require "operating system"-level heavy weapons, yet they can solve the real pain points of small and medium-sized manufacturing enterprises.
The current problem is that no vendor in the domestic Odoo ecosystem has yet seriously taken on this task. Both Kingdee and Yonyou focus on large customers, which may provide a window of opportunity for Odoo partners.
What to watch next
The true outcome of this showdown will not be revealed within 2026. However, there are several signals worth continuing to track:
- Kingdee Lingxi's first major customer signing has been officially announced—whether it can break out from the growth market into large group enterprises is the only criterion for testing whether “AI-native” truly has an advantage;
- Yonyou YonClaw's actual delivery cases——currently still in the invitation-code stage; whether it can provide verifiable ROI data in Q3 2026 will determine the pace of market acceptance;
- National-level AI agent evaluation standards officially released — Industry reaction: Once the standards are implemented, AI operating systems that fail to meet safety specifications will be excluded from procurement by central and state-owned enterprises, creating a hard constraint for both sides.
This article is based on public information from Kingdee's official website (kais2026), Yonyou YonClaw, and industry analysis articles. The specific capabilities of the products are subject to the manufacturers' official releases.
